In order to evaluate changes in GM-CSFR expression, HL-60 cells were incubated in DMSO (1.25% v/v) for 5 to 7 days in order to induce cell maturation. Morphological examination confirmed differentiation o f the cell population from 95% ± 2% immature prom yelocytes to 84% ± 18% morphologically mature cells (myelocytes and metamyelocytes) by day 5 (mean ± SEM o f 3 experiments). Acquisition of the ability to generate superoxide was confirmed using the NBT test which demonstrated that although only 4.5% ± 1.5% were NBT positive at day 0, 56% ± 2.5 % were NBT positive at day 5 (mean ± SEM of 3 experiments). Equilibrium binding studies and Scatchard analysis using ^^^I-GM-CSF have previously shown that uninduced HL-60 cells expressed both low and high affinity GM-CSFR and exposure to DMSO led to a two-fold increase in receptor numbers, without any change in the receptor affinity or the ratio of high to low affinity receptors (Roberts et al, 1994; Table 6.1).
Total RNA was prepared form the HL-60 cells each day after 5 days in DMSO and levels of GM-CSFR a and p chain mRNA were quantified using RNase protection assays. Both a and P chain GM-CSFR mRNA expression
o\ cx> 0) o > Q) c o o c ' c 0) o o Û. 0) o > s z S . 2 Q c g O ' u i < n Q ) </) L _ CL CO X T3 HI (D(/) < (/) z 2 CE 9- E <D 175 150 50 “ 25 D o w n - r e g u l a t i o n U p - r e g u l a t i o n T I T" 5 10 15 “T ” 20 -T - 25 30 —T 35 "T”40 45 TIME (hours)
Figure 6,8 GM-CSFR ct/p mRNA expression during specific ligand-induced down-regulation and subsequent up- regulation;
+ a mRNA, O p mRNA
Down-regulation is achieved by culture in the presence of 360 pM GM-CSF and up-regulation by extensive cell washing (cell wash indicated in figure). Results are expressed as % of control (cells cultured in the presence of 1 IU/1 erythropoietin throughout experiment). Results represent the mean and SEM of 3 separate experiments.
increased during DMSO induced differentiation of HL-60 cells if the signals were standardised for actin mRNA expression (Figure 6.9). If values at day 0 are considered to be 100%, by day 5, expression o f a chain mRNA was 340% ± 6 1 of day 0 values (mean ± SEM of 3 experiments) (Figure 6.9a). p chain mRNA expression increased to 207% ± 69 over the same time period (Figure 6.9b). There is thus a co-ordinate rise in a and P chain mRNAs which correlates with an increase in cell surface protein levels as HL-60 cells differentiate. The ratio of the GM-CSFR a:p mRNA levels (standardised to actin or 18 S ribosomal RNA levels) was 22:1 in uninduced HL-60 cells and 35:1 in day 5 DMSO-induced cells. An actin probe together with a probe for a highly conserved region of the 18S ribosomal RNA gene were used as internal controls, as it has previously been suggested that the actin content of the HL-60 cells increases as they differentiate (Meyer et al, 1983). However, the present data do not confirm a concomitant increase in actin mRNA levels during HL-60 differentiation (Figure 6.9), and therefore actin alone could be satisfactorily used as an internal control. The excess of a chain mRNA observed is in accord with the presence o f dual affinity receptors on these cells.
In TF-1 cells, which also express dual affinity receptors, the ratio o f the a:p chain mRNAs was similar to that o f HL-60 cells. However, in mature neutrophils the ratio was considerably lower at 9:1 (Table 6.2). It is not possible to give accurate absolute quantities on a single cell basis as it cannot be assumed that the mRNA recovery was similar in the different cell types (Section 3.3.3), nevertheless, these data do show that neutrophils have less GM-CSFRa mRNA, relative to GM-CSFRp mRNA, than HL-60 cells which is in accord with the absence of low affinity GM-CSFRs on neutrophils.
DAY 0 1 291 269 Actin 18S n 4 " I
i l
2 3 4 5 -3 #
ActinFigure 6.9 (a) GM-CSFRa mRNA and (b) GM-CSFRp mRNA expression in HL-60 cells undergoing DMSO induced differentiation for 5 days
H L -6 0 TF-1 N e u t r o p h ils Day 0
(uninduced)
Day 5-DMSO induced
(% day 0 uninduced) NA NA GM-CSFRa mRNA 100% 340% ± 61 NA NA GM-CSFRp mRNA 100% 207% ± 69 NA NA Ratio of GM-CSFR a : p mRNA 22 :1 35 :1 50: 1 9 :1
Table 6,2 Expression of GMCSFR mRNA in HL60 and TF-1 cells, and neutrophils (NA = not applicable)
6.4 Discussion
The expression of GM-CSFRs has been extensively studied in a range o f haematological cell types (Budel et al, 1993; Khwaja et al, 1993; Roberts et al, 1994) but there is little information about the molecular regulation o f these receptors. The work presented in this chapter was performed in order to define the changes in GM-CSFRa and p chain mRNAs that occur during differentiation and in response to dynamic changes in cell surface receptor expression.
The half life of the GM-CSFR protein in TF-1 cells grown in steady state conditions in the presence of Epo and the translation inhibitor cycloheximide, was approximately 4 hours. This is considerably shorter than has been reported for other receptors, for example, up-regulated insulin receptors in dexamethasone treated mouse fibroblasts have a receptor half-life of 19 hours (Knutson, 1992). The rapid down-regulation of the receptors under these conditions was unlikely to be due to ligand induced internalisation by contaminating GM-CSF in the extracellular medium for several reasons. The cells had been washed extensively prior to transfer to Epo and the estimated GM-CSF contamination in the Epo culture phase is in the low femtomolar range. In addition, experiments utilising TCA precipitation demonstrated that internalised GM-CSF is degraded, and would not therefore be released into the supernatant. In accord with this, if the washed cells were cultured without Epo there was cell death by apoptosis detectable by 36 hours (data not shown). This suggests that loss o f receptors from the cell surface could either be by enzymatic cleavage as demonstrated for the M-CSFR (Downinggr al, 1989) or by internalisation occurring even in the absence of specific ligand as in the insulin receptor (Knutson, 1992). This latter process appears to be metabolically controlled and is the rate-limiting step in the expression of the insulin receptor, whereas receptor recycling to the cell surface does not appear to be a regulated process. Similarly, internalisation o f the
epidermal growth factor receptor (EGFR) in the absence o f ligand has been demonstrated in the epidermoid carcinoma cell A431 and skin fibroblasts to have a receptor half life of 9 - 16 hours (Krupper al 1982). Despite this apparent GM- CSFR turnover, surface expression of the receptor remained constant over 6 hours, even in the absence of transcription as shown by GM-CSF binding in the presence of actinomycin D. As the half-life of the GM-CSFRa chain mRNA was found to be between 1 and 3 hours (Figure 6.2a), this indicated that there was a fall in a chain mRNA without a concomitant decrease in l^^I-GM-CSF binding. This could either be due to more efficient a chain translation, or to a change in GM-CSFR affinity, possibly as a result of a change in relative numbers o f a and P chains leading to alternative oligomerisation of these chains (Budel et al, 1993; Wheadon et al, 1995). However, there was no change in the surface expression o f either chain during this period, as measured by surface binding o f GM- CSFRa and P chains with monoclonal antibodies. Therefore the level of a chain protein was maintained at the cell surface, even when the a chain mRNA had fallen to 25%. This suggests that translational control of the GM-CSFRa chain production can occur to maintain steady-state receptor levels.
The surface expression o f the GM-CSFR can be rapidly modulated by the presence o f specific ligand, with nearly all high affinity receptors and approximately 70% of low affinity receptors being down-regulated by 4 hours. If the GM-CSF was then removed by extensive washing, the surface ligand binding ability is up-regulated to the pre-GM-CSF exposure levels after 18 hours. For ligand-binding to increase to steady state levels, one or more o f four conditions must be operative: receptors internalised in the presence of GM-CSF must be recycled to the cell surface; the steady state turnover o f the receptors must be markedly reduced; the affinity of receptors present must increase; or the rate of receptor chain synthesis must be increased relative to the steady state level. The present data demonstrates no evidence for significant receptor recycling. Firstly,
receptor up-regulation was inhibited in the presence of cycloheximide suggesting that this process requires new protein synthesis and that a pool o f intracellular receptors is not available for surface expression. Secondly, once internalised, the receptor bound ligand was degraded, as demonstrated by an increase in radioactivity in the TCA-soluble fraction containing low molecular weight fragments of ^^^I-GM-CSF. Inhibition of lysosomal acidification by incubation with 10 mM ammonium chloride prevents ligand dissociation and degradation of the ligand, but did not lead to an alteration either in GM-CSFR re-expression after ligand-induced down-regulation (as measured by binding of radioligand) or re-expression o f a and P chains (as measured by specific monoclonal antibodies). This suggests that recycling o f the receptor from the dissociated receptor-ligand complex does not make a major contribution to its up-regulation. In addition, although Knutson (1992) demonstrated that changes in the internalisation rate constant of the insulin receptor can lead to changes in the relative number of recycled to degraded receptors, we could find no alteration in the rate of ligand-independent receptor loss from the cell surface which might explain an increase in receptor expression (Figure 6.7). Using the murine myelomonocytic line WEHI-3BD+, Walker & Burgess (1987) have reported that a considerable proportion of the murine GM-CSFR is recycled to the surface, and although new synthesis does contribute to the re-expression o f receptors, this occurs in the first hour after down-regulation. In their model cycloheximide alone did not inhibit receptor re-expression after ligand-induced down-regulation and only a combination of 50 mM ammonium chloride and cycloheximide resulted in inhibition o f receptor re-expression. The differences between the results presented here may be due to differences between species or between the different cell models used.
The internalisation and intracellular degradation of GM-CSF and its receptor would therefore be a mechanism by which the effects of the ligand are down-
regulated and are in agreement with the findings o f Khwaja et al, 1990. This has also been well described for other growth factor receptors EGFR (Felder et at, 1990) and M-CSFR (Downing et al, 1989). Both these receptors homo-dimerize on interaction with their cognate ligands, whereas interaction of GM-CSF with its receptor results in a heterodimer or complex o f a and P chains. The data presented here would suggest that both the a and P chains are degraded on internalisation of the receptor complex. Inhibition of receptor up-regulation by cycloheximide indicates that the major mechanism for up-regulation is by increased protein synthesis. However, throughout the period of receptor up- regulation, a and p mRNA levels remained constant (Figure 6.8) suggesting that the increase in receptor synthesis is due to increased translational efficiency. Translational control has been described for a number of proteins including ferritin, vertebrate ribosomal proteins, Xenopus oocyte maturation and the yeast GCN4 system (Melefors & Hentze, 1993).
During DMSO-induced maturation of HL-60 cells, up-regulation of both high and low affinity receptors was observed, indicating an increase in both GM- CSFRa and p chain production. This was associated with a commensurate rise in a and P mRNA levels, suggesting either increased transcriptional activity, or an increase in mRNA stability associated with differentiation. One way o f confirming transcriptional regulation would be to perform nuclear run-off assays, using nuclear extract o f induced HL-60 cells. Nuclear run-off assays are, however, notoriously unreliable. When these assays were performed using the murine IL-3 receptor AIC2A, nascent transcripts were barely detectable in the MC9 cell line (Hara & Miyajima, 1994). In mature neutrophils only a single class of the high/intermediate affinity receptors was seen, indicating a relative increase in the surface p chains compared to a chains. The fact that induced HL-60 cells do not have this expression pattern is in line with previous reports that HL-60 cells do not acquire all the characteristics of mature neutrophils, in that they lack
specific granules (Newburger et al, 1979) and have monocyte DNase (Roberts et al, 1990b). Analysis o f the relative amounts o f a and p chain mRNAs in neutrophils indicated a lower GM-CSFRa to P mRNA ratio compared to HL-60 cells, which is in accord with the expression of a and P chain proteins.
In conclusion, the expression o f the GM-CSFRs is highly regulated follow ing both ligand-induced down-regulation o f receptors and during differentiation but the mechanism of regulation appears to differ. Whereas changes GM-CSFR protein during differentiation are associated with co-ordinate changes in mRNA levels, rapid changes in receptor expression induced by transient exposure to GM-CSF do not involve any change in mRNA levels and are at the level of translation.
— CHAPTER 7 —
SUMMARY
Summary
Early studies with murine and human bone marrow demonstrated the presence of a relatively low number of GM-CSF specific surface receptors. The cloning o f the GM-CSFRa chain and the common P chain has enabled the study o f the receptor at the level of mRNA, as well as cell surface protein expression. The problems associated with the quantification of biological systems in general, and molecular cell events in particular, have resulted in the majority of studies to be qualitative in nature. The data presented in this thesis attempts a quantitative approach of molecular events utilising low substrate concentration.
Chapter 3 highlights the problem of attempting to measure mRNA